D-01
Draw first.
If it cannot be specified, it cannot be trusted. Every system I ship starts as a document that someone else could build without me in the room.
§1.0 IDENTITY · DRAWING OF A PERSON
Shreyas B S
Agentic Systems Engineer
Systems that explain themselves.
I build AI systems the way engineers build bridges: drawn completely before they are built, instrumented so they can be trusted, and shipped to survive production. My work sits where multi-agent systems meet evaluation, making autonomous systems provable instead of plausible.
currently: AI intern · Formi (Agentic Universe) · CSE @ MIT Manipal, class of 2028 · Bengaluru
- India AI Impact Buildathon 2026
- Top 15 / 42,000+
- IISc Arbitrage Arena 2026
- 2nd Runners Up / 420+
- Patent filings
- 4
- Technoxian World Cup
- 5th / 80+
§2.0 DOCTRINE
How I think
Five rules govern everything I ship. They are not aspirations; each one is enforced somewhere in my work, and most of them are enforced on this very page.
D-02
Determinism is respect.
Same seed, same behavior. A system that cannot be replayed cannot be debugged, audited, or believed. This site contains zero calls to Math.random(), and that is enforced at build time.
D-03
Errors are asymmetric.
Every failure mode has a price tag, and the prices are never equal. In voice AI, a false interruption is catastrophic and a missed one is recoverable, so you default to silence and escalate monotonically. Find the expensive failure and design for it.
D-04
Evals before features.
A capability you cannot measure is a liability you have not priced. I wrote 54 deterministic metrics for a voice pipeline before anyone asked for a dashboard.
D-05
Trust is infrastructure.
Reputation, provenance, and auditability are not add-ons. They are the load-bearing walls of agent systems, which is why my patent filings and my IEEE submission are all about making agents accountable.
§3.0 TRAJECTORY
The line so far
-
2024
MIT Manipal
Started Computer Science Engineering, class of 2028. Joined RoboManipal, the university's robotics team, and found the lab I would not leave.
-
2025
First production agents
Joined VersionTwo as an Agentic AI Developer in June. Designed a 10-agent content workflow that cut creation time by 80% for DeepTech founders. FarmBot work at RoboManipal began winning: Technoxian World Cup, 5th of 80+.
-
EARLY 2026
The competition season
2nd Runners Up of 420+ teams at the IISc Arbitrage Arena. Top 15 of 42,000+ entries at the India AI Impact Buildathon with a multi-agent scam-baiting honeypot.
-
2026
Production and protection
AI intern at Formi (Agentic Universe), building evaluation infrastructure for live voice agents. Four patent filings in development, and the karma reputation paper submitted to IEEE SSRR 2026.
-
NEXT
The open segment
The line continues below in §4.
§4.0 SIGNAL
What I'm optimizing for
I want to work on agentic systems that have to survive production: evaluation infrastructure, safety tooling, voice agents, multi-agent orchestration. If you are building in that space and the hard parts are still hard, I want to hear about them.
open to: internships research collaborations agent infrastructure
off the clock
Robotics lab nights at RoboManipal. Teaching 60+ juniors to build agents from raw APIs, no frameworks allowed. And an unhealthy fascination with why systems fail rather than how they work.
§5.0 TRANSMISSION
Start a conversation
> spawn_conversation()
resume --format pdf pending upload
This was the drawing. The machine room is on the other side.
[§1] thesis.load()
Systems that explain themselves.
I'm Shreyas. Right now my systems are evaluating live enterprise calls, baiting scam callers, and grading themselves in production. This site runs the same way.
currently: AI intern · Formi (Agentic Universe) · CSE @ MIT Manipal, class of 2028 · Bengaluru
- India AI Impact Buildathon 2026
- Top 15 / 42,000+
- IISc Arbitrage Arena 2026
- 2nd Runners Up / 420+
- Patent filings
- 4
- Technoxian World Cup
- 5th / 80+
[§2] experience.mount()
Experience
the pipeline, drawn: one call through four stations
80% content creation time reduction for DeepTech founders
Technoxian World Cup 5th / 80+ · Farm Robotics Challenge USA 2026
Teaching & mentoring
TDA Gen AI & Agentic AI Bootcamp · RoboManipal
I run the Gen AI and Agentic AI Bootcamp at MIT Manipal, pure API and Python, no frameworks, and mentor 60+ juniors at RoboManipal. 60+ people now build and ship agent systems who didn't before. One node, fanned out.
[§3] machines.mount()
Projects
The triage engine is rules-first: "why this band?" explanations cite the exact protocol rules that fired, from ANC danger signs to snakebite response. AI is routed by cost and stakes: small model for extraction, large model only for high-risk referral language.
ANC · IMNCI · HBNC · PPH · TB DOTS · snakebite · NCD PEN · IDSP
illustrative model · competition dataset pending owner
2nd Runners Up / 420+ · IISc Arbitrage Arena 2026
Most FX momentum is just the dollar moving. Removing that common factor with PCA leaves residuals that mean-revert, and a cross-sectional momentum signal on those residuals is a cleaner trade than the raw pairs.
common factor out · residual signal in
The research behind the temporal-GNN paper in §4: build multiplex networks over India's top 50 companies, then learn how negative news and volatility shocks travel through them.
code for the §4 temporal-GNN paper
The same thesis as everything else here: decisions you can audit beat decisions you can only trust. Every classification cites its evidence.
every decision → span → token → rule
review intervals grow as recall strengthens: 1d · 3d · 7d · 16d · 35d
[§4] research.verify()
Research & Patents
- Karma-based reputation economies for heterogeneous human-AI-robot systems. IEEE SSRR 2026.
- IoT water quality monitoring, lean 5HO. India International WASH Conference 2026.
- Temporal GNN analysis of financial networks.
- MCDM siting of solar parks.
Patent filings
Status: in development
Patent filing
MTSRD
Multi-turn semantic reconstruction detection. An attacker can assemble a harmful request piecemeal across many innocent turns; MTSRD slides a window over the conversation and asks, for each span, whether the fragments recombine into a known harmful intent. Detection happens at recombination time, not at the single-turn level where each fragment looks clean.
Status: in development
Karma Protocol
A reputation economy for AI agents. Every agent action earns or burns karma against a scored ledger:
K = 0.35·Safety + 0.30·Accuracy + 0.20·Consistency + 0.15·Performance decay: 3% / day
Underpins the reputation-economy paper submitted to IEEE SSRR 2026 above.
Status: in development
Patent filing
Slip Estimation
Traction-aware locomotion for field robots, out of the RoboManipal lineage. Title-level disclosure only.
Status: in development
Patent filing
Attentive Fuzzy Logic
Attention-weighted fuzzy control, RoboManipal lineage. Title-level disclosure only.
Title pending owner confirmation
Fourth filing
A fourth patent filing exists; its public-safe title is pending confirmation before publication here.
[§5] schematic.energize()
One system
All subsystems report to the same core. Watch the signals: multi-agent systems, evaluation, trust.
Systems that explain themselves.
> spawn_conversation()
resume --format pdf pending upload
That was the machine room. The person who drew it is on the other side.